Studying the genetics of Alzheimer's Disease (AD) has added significantly to our understanding of the disease. During the last six years it has been established that familial early onset Alzheimer's disease (FAD) is a genetically heterogenous disease that can be caused by mutations in at least three different genes: the beta-amyloid protein precursor (APP) gene on chromosome 21, the presenilin 1 (PS-1) gene on chromosome 14 and the presenilin 2 (PS-2) gene on chromosome 1 (1-3). Since other families exist that do not carry mutations within any of these genes it is very likely that there are other as yet unidentified FAD genes. In vitro experiments suggest that mutations in each of the known genes cause AD through changes in APP processing that lead to elevated levels of total Abeta or specifically increase Abeta42 (4). This provides strong support for the "Amyloid Hypothesis" of AD pathogenesis. The study of the genetics of late onset AD has also led to the identification of the first genetic risk factor for AD. The epsilon 4 allele of the apolipoprotein E (ApoE) gene has been shown to increase risk for AD in every population studied although the magnitude of the increase in risk varies between populations. Although it is still uncertain how the ApoE4 allele increases risk for AD there is some evidence to support the hypothesis that ApoE4 modifies amyloid deposition by an unknown mechanism. However, since there are many individuals with AD who have no ApoE4 alleles there must be other risk factors for late onset AD. We hypothesize that at least some of these risk factors are genetic and that they may also modify amyloid deposition. The aim of this proposal is to use a genetic linkage strategy to identify new genetic risk factors for late onset AD. As our test sample we will use three hundred caucasian sib pairs with an age of onset of AD over the age of sixty five years to perform a 20cM genome-wide screen using microsatellite markers. Genomic regions showing evidence of linkage will be followed up with flanking markers in the same sample and in a second sample of equivalent size also selected on the basis of racial origin and age of onset. It is anticipated that by restricting the racial origin and age of onset of our initial samples we will reduce the likely genetic heterogeneity and increase our chances of detecting a second risk factor. Candidate genes in genomic regions that continue to show evidence of linkage will then be followed up using a case control association approach in caucasians and in other ethnic groups. Finally, we will complete a 20cM genome screen in the replication sample increasing the effective sample size to 600 sib pairs, enabling us to look for genes of smaller effect size.